ITP: a historical perspective


Dr Roberto Stasi, Department of Haematology, St George’s Hospital, Blackshaw Road, SW17 0QT, London, UK.


A clinical syndrome of bleeding and purpura consistent with a diagnosis of immune thrombocytopenia (ITP) was described by Werlhof long before platelets were identified as the cellular component of blood playing an essential role in primary haemostasis. Although a role for the spleen was suggested nearly a century ago, the pathophysiology of ITP has remained elusive for many decades. During this time Werlhof’s disease was renamed idiopathic thrombocytopenic purpura, from which the acronym ITP originally derives. The second half of the 20th century brought recognition of the autoimmune components of ITP, and hence the need for a new standard nomenclature, which has recently been accepted. ITP currently stands for Immune Thrombocytopenia, a name that more appropriately reflects the low platelet count rather than purpura as the main feature of the disease, as well as to defining its underlying nature. Advances in our knowledge of the disease have paralleled the availability of new therapeutic agents, and we are now entering an era of pathophysiologically-based treatment options.

Immune thrombocytopenia is an acronym for primary immune thrombocytopenia, previously referred to as idiopathic thrombocytopenic purpura. ITP is an acquired autoimmune disorder characterized by isolated thrombocytopenia in the absence of conditions known to cause thrombocytopenia, such as infections, other autoimmune disorders, drugs, malignancy, etc. In up to 40% of adults it may have no clinical manifestations at all and in the majority of cases it causes only localized haemorrhaging in skin or mucous membranes that are usually of little to no clinical consequence (petechiae, purpura, ecchymoses, epistaxis); more rarely, it can be associated with severe bleeding events such as intracranial haemorrhage (ICH). Most ITP patients are asymptomatic in the presence of platelet counts greater than 50 × 109/l (Stasi et al, 2008a).

The story of ITP is intriguing, as advances in our knowledge of the disease reflect those in many other fields of medicine, from molecular biology to immunology, from drug therapy to surgery.

The pathophysiology and the management of ITP have been the subject of other recent reviews published in this journal (Provan & Newland, 2002; Cooper & Bussel, 2006; Psaila & Bussel, 2008; Nugent et al, 2009; Molineux & Newland, 2010). The present review will outline landmark observations and discoveries in the history of this disease.

Early descriptions of purpura

Before platelets were identified, the identification of ITP was based exclusively on the presence of purpura in an otherwise healthy individual. Purpura is a Latin word meaning ‘purple’. It derives from the Greek πορϕύρα (porphyra), a word of Semitic origin, denoting a Mediterranean sea snail (shellfish) of the Muricidae, or murex, family from which a reddish dye was obtained in ancient times.

In medicine, purpura is a general term for reddish-purple skin lesions produced by bleeding in the dermis or subcutaneous tissues. It was recognized as a clinical symptom as early as the Greco-Roman period by physicians such as Hippocrates and Galen, who described the condition as red ‘eminences’ or spots associated with pestilential fevers (Rolleston, 1934). Since then, from time to time various attempts have been made to subdivide purpura according to various characteristics. However, the lack of specific clinical correlates made these classifications of little or no utility in the differential diagnosis of diseases.

It was not until 1025, in the monumental 14-volume treatise The Canon of Medicine, that the Persian physician and philosopher Abu Ali al-Hussain Ibn Abdallah Ibn Sina (c. 980–1037), more commonly known by his Latinized name Avicenna, briefly described a chronic purpura that could fit the diagnosis of ITP (Jones & Tocantins, 1933).

By the XVI century, the term purpura was used in relation to infectious diseases, such as typhoid fever, but was also used to describe other conditions often referred to as ‘purpura sine fever’ or ‘petechia sine fever’.

In 1556, the Portuguese physician Amatus Lusitanus (1511–1568) published accounts of his cases and their treatment (Fig 1). In his book, entitled Curationum Medicinalium Centuriae, he described a boy who developed dark macules and bloody discharges for several days in the absence of fever (Morbus Pulicaris Absque Febre), and eventually recovered spontaneously (Lusitanus, 1556).

Figure 1.

 Left panel. Title page of one of Lusitano’s Centuriae. Lusitano wrote seven Centuriae, each containing descriptions of hundreds of different case histories that he had treated, with commentaries and discussions. Right panel. Statue of Amato Lusitano, born João Rodrigues, in his Portuguese hometown Castelo Branco. These images are in the public domain.

In 1658 Lazarus de la Riviere (Riverius), physician to the King of France, described purpura as ‘Purple specks or spots like flea-bitings… which the Italian physicians call peticulae or petechiae…’ (Riverius, 1668). He suggested that purpura was a manifestation of a systemic bleeding disorder, something he called ‘thinness of the blood’.

In 1735, the German physician and poet Paul Gottlieb Werlhof (1699–1767) provided the classic clinical description of ITP, calling it ‘morbus maculosus haemorrhagicus’ (Fig 2). He reported the case of a 16-year-old girl with cutaneous and mucosal bleeding that occurred after an infectious disease (Werlhof, 1735). The girl bled from her nose and mouth and vomited ‘very thick, extremely black blood. Immediately there appeared about the neck and on the arms, spots partly black, partly violaceous or purple…’. This condition was subsequently named after Werlhof, and the eponym is sometimes still used instead of ITP.

Figure 2.

 Left panel. Title page of Werlohf’s book ‘De Variolis et Anthracibus’. Right panel. Paul Gottlieb Werlhof (1699–1767). A painting by Dominicus van der Smissen, 1740. Photographic reproduction. Werlhof was a native of Helmstedt, where he also completed his studies. In addition to becoming one of the most influential physicians in Europe, he was also an accomplished poet. While his medical treatises were written in Latin, the ‘official’ scientific language of his times, Werlhof composed his poems and hymns in German. These images are in the public domain.

In his book On Cutaneous Disease, published in 1808, the English dermatologist Robert Willan (1757–1812) described many forms of skin disease (Willan, 1808). Elaborating on the earlier studies by Riverius and Werlhof, Willan assigned to the ancient term purpura the meaning it retains to this day and classified purpura as simplex, haemorrhagic, urticarial, and contagious. Willan noted that purpura simplex (the type most consistent with ITP) ‘…chiefly affects women and children…’ and is characterized ‘…by an appearance of petechiae without much disorder of the constitution. The petechiae are diffused chiefly over the arms, legs, breasts and abdomen. They are largest on the legs, though seldom confluent’.

In 1899 Henoch in Berlin differentiated purpura simplex with bleeding signs of the skin only (today known as dry purpura) from purpura haemorrhagica with mucosal bleeding (today known as wet purpura) (Henoch, 1899).

The discovery of platelets and their link with ITP

It would have been impossible to visualize platelets before 1826, when the problem of chromatic aberration in the compound microscope was overcome by Joseph Jackson Lister (1786–1869). Lister devised a method of lens combination, by which the aberrations of one lens could be neutralized by a second. With Lister’s microscope it was possible to resolve objects about 1 μm apart (Robb-Smith, 1967).

In 1841 two English physicians, George Gulliver (1804–1882) and William Addison (1803–1881), independently made the first drawings of platelets (Robb-Smith, 1967). However, only Addison associated these platelets with fibrin formation. He noted that the ‘liquor sanguinis contained a great number of extremely minute molecules or granules, varying in size, the largest being at least eight or ten times less than the colourless corpuscles… Whilst examining these minute bodies, I observed the coagulation of the fibrin commence’. Addison did not believe they were independent cellular elements, but considered these particles as granules discharged from ruptured leucocytes.

In 1842 the French physician Alfred Donné (1801–1878) described three types of particles in blood, of which the erythrocytes and leucocytes were clearly identified. The third cellular component, probably platelets, was more elusive and described as ‘les globulins du chyle’, as he thought that these corpuscles were the product of chyle.

Around the same time, in Germany, Franz Simon and Gustav Zimmermann independently observed that blood rendered incoagulable by potassium ferrocyanide contains ‘billions’ of small colourless bodies. Simon thought that they were molecules of fibrin which were red blood cell precursors, while Zimmermann felt that they were ‘elementary bodies’ which came from the lymphatics and speculated that they were an early stage of maturation of erythrocytes.

In 1865 Max Schultze (1825–1874) published the first accurate description of platelets as a part of a study devoted to white blood cells (Schultze, 1865). According to his description, blood of healthy individuals contains irregular aggregates consisting of small, colourless spherules or granules for which he proposed the term granular masses. With the over-simplified attitude typical of his time, Schultze inferred they had to be considered degenerate and disintegrated leucocytes. The concept of Schultze’s corpuscles as ‘dust of the blood’ originating from leucocytes mainly during infectious diseases was supported by many other contemporary haematologists.

In 1874 William Osler (1849–1919), a young Canadian physician, gave a description of platelets as distinct cellular elements that are a normal constituent of circulating blood (Fig 3). He thought that these particles were not artefacts or the by-product of leucocyte degeneration, but could not speculate about their nature (Osler, 1874).

Figure 3.

 Left panel. William Osler delivered many lectures before learned societies, such as the Cartwright Lectures before the Association of Physicians and Surgeons in New York City in 1886. Right panel. Sir William Osler seated at desk. Photographic reproduction. Public domain. National Library of Medicine (NLM).

The standstill in progress ended suddenly, in 1879, when Paul Ehrlich (1854–1915) published his technique for staining blood films and his method for differential blood cell analysis (Ehrlich, 1879).

The term ‘platelets’ (blutplättchen) was introduced in 1882 by the Italian pathologist Bizzozero, (1882). By means of in vivo and in vitro investigations, Bizzozero demonstrated that platelets were an independent cell line with the specialized function of haemostasis and brilliantly noted that haemostasis and blood coagulation were not synonymous (Fig 4). However, despite his acumen he could not demonstrate the origin of platelets.

Figure 4.

 Left panel. Bizzozero’s article published in German in 1882 in which he clearly described platelets as a third morphological element of the blood, unrelated to erythrocytes and leucocytes, and their role in in thrombosis and haemostasis. Right panel. Giulio Bizzozero (1846–1901), photographic reproduction. Bizzozero was born in Varese, not far from Milan, and graduated with honours from Pavia University Medical School at the age of 20. Shortly afterwards he began his academic career in Pavia as a deputy professor of general pathology and lecturer of histology. In 1872, Bizzozero moved to the University of Turin where he had been appointed professor of General Pathology. He remained in Turin until his death in 1901, of pneumonia. These images are in the public domain.

The link between Willan’s purpura simplex and thrombocytopenia was apparently first reported by Brohm in 1883 and published in the dissertation of Krauss (Brohm, 1883). In 1887 the Belgian histologist Denys, called attention to the fact that platelets were diminished during the active phase of the purpura and increased when the haemorrhages ceased (Denys, 1887). Two years later Georges Hayem (1841–1933) in Paris was able to confirm this observation and at the same time he also noted that the remaining platelets (which he termed haematoblasts, believing they were primitive erythrocytes) were large and did not support clot retraction normally (Hayem, 1889).

In 1890, in Baltimore, William Henry Howell (1860–1945) first described megakaryocytes (Howell, 1890) and in 1906, in Boston, the pathologist James Homer Wright (1869–1928) conducted studies of feline bone marrow that led him to postulate that platelets were fragments of megakaryocytes (Wright, 1906).

In 1912 William Waddell Duke (1883–1945) called attention to the fact that the bleeding time may be greatly prolonged while the clotting time is normal (Duke, 1912). This prolongation of bleeding time after a pin prick may in some cases exceed an hour, whereas normally, bleeding ceases within 3 min.

Evolving concepts in the pathophysiology of ITP

Early theories

The mechanisms responsible for the thrombocytopenia in Werlhof’s disease soon became a matter of controversy.

Immune-mediated mechanisms were first suggested in 1905 by Marino, who produced an antiplatelet antibody by injecting rabbit platelets into guinea pigs (Marino, 1905). A few years later this theory received further support by the work of Ledingham, who caused experimental purpura in guinea pigs after inoculation with an antiserum prepared by immunizing rabbits with guinea pig platelets (Ledingham, 1914).

In 1915 the German physician Ernest Frank postulated a marked diminution in platelet production by megakaryocytes. He proposed that ‘Die Essentielle Thrombopenie’ (as he renamed Werlhof’s disease) resulted from toxic suppression of the megakaryocyte by a substance produced in the spleen (Frank, 1915).

In 1916 a medical student in Prague, Paul Kaznelson (Fig 5), challenged Frank’s idea and proposed that, in analogy with haemolytic anaemia, essential thrombocytopenia resulted from increased platelet destruction in the spleen. Kaznelson convinced his tutor, Professor Doktor Schloffer (Fig 5), to perform a splenectomy in a 36-year-old woman with a history consistent with our current definition of chronic ITP. The patient for most of her life had suffered from easy bruising, frequent epistaxis and heavy menstrual bleeding. The platelet count was 2 × 109/l prior to splenectomy and rose to 500 × 109/l within 4 weeks from surgery with complete resolution of the purpura (Kaznelson, 1916). However, in the next two cases only temporary increases of the platelet count were observed (Kaznelson, 1919).

Figure 5.

 Left panel. Paul Kaznelson (1898–1959) appearing in his application for UK nationality in 1950. Kaznelson was born in Warsaw, then a part of the Russian Empire, from a family of Jewish origin. He acquired a Czechoslovakian nationality by naturalization in 1919, but had to leave the country to escape the Nazis in 1939. He managed to land in Liverpool in 1940 and remained in the UK throughout World War 2, working in many British hospitals. Source: Yoshida Y. Historical review. The light and shadow of Paul Kaznelson: his life and contribution to hematology. Annals of Hematology 2008;87:877–879. The image is licensed under the Creative Commons Attribution-Share Alike 2.5 Generic license ( Right panel. Hermann Schloffer (1868–1937). Schloffer was a native of Graz, in Austria. He spent several years in Prague and was a professor at Charles University in Prague. He was Kaznelson’s tutor and the surgeon who actually first performed a therapeutic splenectomy in a patient with ITP.Source: Medizinische Universität Innsbruck ( This image is in the public domain.

A third short-lived theory, not based on experimental evidence and rebutting the conclusions of previous observations, proposed that the diminution of platelets in the peripheral circulation was not the cause but the result of haemorrhage (Tidy, 1928). According to this theory, platelets are withdrawn from the circulation and fixed in the bleeding areas as a defence mechanism.

In 1938, Troland & Lee extracted a substance (‘thrombocytopen’) from spleens of patients with chronic ITP and injected it into rabbits (Troland & Lee, 1938). Thrombocytopen consistently produced a rapid although transient fall of the platelet count. This was the first evidence of the involvement of a transmissible plasma factor in ITP. The development of thrombocytopenia in experimental animals following injection of splenic extracts from patients with ITP was subsequently confirmed by other investigators (Hobson & Witts, 1940; Otanasek & Lee, 1941; Rose & Boyer, 1941; Paul, 1942; Uihlein, 1942; Cronkite, 1944).

In a seminal paper published in 1946, Dameshek and Miller reported key findings for our current interpretation of the mechanisms of thrombocytopenia in ITP (Dameshek & Miller, 1946). They observed that the numbers of megakaryocytes and platelets were correlated across a variety of disorders, including leukaemia and pernicious anaemia, and that the platelet count was consistently low when megakaryocytes were reduced in number. However, while platelet numbers were decreased in ITP, megakaryocyte numbers were normal or increased, but only a third or fewer of megakaryocytes showed evidence of platelet production. In addition, they also found that the numbers of a larger intermediate megakaryocyte, the promegakaryocyte, were decreased in patients with ITP. They concluded that the decrease in blood platelets was the result of a severe reduction in platelet production by megakaryocytes. Interestingly, effective platelet production appeared to increase after splenectomy. This observation led these authors to conclude that the disease was due to a fundamental abnormality of the spleen which ‘exerts an unusual effect upon the production of platelets from the megakaryocytes in the marrow’.

Harrington–Hollingsworth experiment

The debate about the mechanisms of the thrombocytopenia in ITP, peripheral destruction of platelets versus impaired platelet production, was apparently settled in 1951 by the experiment undertaken by William J. Harrington and James W. Hollingsworth, two haematology fellows working at Barnes-Jewish Hospital in St. Louis, Missouri (Harrington et al, 1951). In his historical review of ITP, Paul Imbach reports that Harrington et al had observed a child with purpura born to a mother with chronic ITP, a condition already described by Dohrn in 1873 (Imbach et al, 2002). Purpura in the child resolved 3 weeks later, although the mother still had ITP. It was therefore possible that a humoral antiplatelet factor had been passed from mother to child. To test this hypothesis, a middle-aged woman with persistent thrombocytopenia postsplenectomy who had been admitted with bleeding was asked to participate in the experiment. Serology found that her blood group was O-positive. This was the same as Harrington’s blood group, so Harrington was to be the first volunteer. He organized an exchange transfusion of approximately 500 ml of blood between the patient and himself. Prior to the exchange transfusion the patient’s platelet count was measured at 5 × 109/l and Harrington’s was measured at about 250 × 109/l. Harrington also underwent a bone marrow aspirate to demonstrate normal megakaryopoiesis. The exchange transfusion itself proceeded without complication. However, shortly after the transfusion when he was about to undergo a repeat bone marrow aspirate, Harrington experienced a generalized seizure. He recovered spontaneously from this after a few minutes (Altman, 1998). Post-transfusion platelet count confirmed the researchers’ hypothesis. The patient’s platelet count was not significantly different at 6 × 109/l, whereas Harrington’s had fallen dramatically to 10 × 109/l. Over the following 3 d, Harrington, who in the meantime had been admitted at Barnes Hospital, experienced a series of complications related to progressive thrombocytopenia, including mucosal and cutaneous petechiae, gingival bleeding, epistaxis and rectal bleeding. In an effort to avoid ICH, Harrington slept upright, supported by pillows, as a means for reducing intracerebral pressure. Within a week, Harrington’s platelet count had recovered and he was able to return to normal work. He performed similar experiment on volunteers (healthy faculty staff and patients with terminal cancer), confirming his original finding (Fig 6), and concluded that some factor in the plasma of ITP patients was destroying the platelets. Bone-marrow investigations performed on himself and some of the other volunteers showed an increased number of megakaryocytes during the thrombocytopenia. He later provided evidence that the plasma factor resided in the gamma-globulin fraction. Of additional interest, the blood plasma fraction, collected from two of the 10 patients with ITP after their platelet levels had returned to normal following splenectomy, elicited the same degree of thrombocytopenia in the volunteers as had their blood plasma fraction before they underwent splenectomy. The effect of splenectomy on the patients was mixed, ranging from remission to no rise in the platelet count.

Figure 6.

 Left panel. The graph shows the results of the Harrington–Hollingsworth experiment. There was a rapid development of thrombocytopenia, followed by a return to normal platelet levels, in healthy volunteers who received plasma from patients with ITP. The lowest result was Harrington himself (from Harrington et al, 1951). Right panel. William J. Harrington, Sr (1923–1992). Harrington was born in Norwood, Massachusetts, and graduated from Tufts University Medical School in 1947. After conducting his self-experiments, he continued his research at Washington University and in 1954 became the director of its division of haematology. In 1964 he moved to the University of Miami School of Medicine, where he founded the Center for Blood Diseases. The Blood Center was later named after him. These images are in the public domain.

The Harrington and Hollingsworth experiment had unequivocally demonstrated that ITP is characterized by reduced platelet survival due to a humoral factor. That same year Evans et al (1951) suggested that the thrombocytopenic factor was an anti-platelet antibody. However, data from other lines of research were not entirely consistent with a model of ITP based only on platelet destruction.

Antiplatelet autoantibodies and their role in the pathogenesis of ITP

Phase contrast microscopy studies by Pisciotta et al (1953) confirmed the findings of Dameshek and Miller, describing several abnormalities of megakaryocytes in ITP. They hypothesized that, in ITP, megakaryocytes suffer from a defect in maturation as well as from abnormalities in platelet formation and release and that a powerful platelet agglutinin present in ITP was equally capable of attacking megakaryocytes and the platelets surrounding and budding from megakaryocytes.

The research of Shulman et al (1965), further delineated that the thrombocytopenic factor in the plasma of ITP patients was associated with immunoglobulin G (IgG), and that the degree of thrombocytopenia in healthy subjects following infusion of ITP plasma depended on the dose of plasma infused.

Studies performed in the 1970s demonstrated that platelet-associated IgG was increased in c. 90% of chronic ITP patients, stimulating new investigations of the antibody component of ITP (Dixon et al, 1975). The nature of the IgG producing thrombocytopenia was eventually unravelled in 1982 by the experiments of van Leeuwen et al (1982). They noted that sera or eluates of platelets from patients with ITP would, in each case, bind to normal platelets but only about one quarter would bind to the platelets of patients with Glanzmann thrombasthenia. They speculated that ITP patients produced autoantibodies against either platelet glycoprotein (GP) IIb or GPIIIa because thrombasthenic patients lack these proteins. Since that time, several laboratories have provided direct evidence for the presence of autoantibodies against GPIIb/IIIa and other platelet antigens in ITP (Woods et al, 1984a,b; Kiefel et al, 1987; McMillan et al, 1987).

In the meantime, platelet kinetic studies using radiolabelling methods had reignited the debate about the mechanisms of thrombocytopenia in ITP. Early investigations using 59Cr in patients with severe thrombocytopenia were technically challenging and could only be performed with large numbers of heterologous platelets. These studies showed shortened platelet survival and high rates of platelet production, ranging from four to nine times normal in patients with ITP (Harker & Finch, 1969; Harker, 1970; Branehog et al, 1974). With the introduction of 111In, a radioisotope with characteristics more suitable for the study of platelet kinetics and distribution, the small numbers of autologous platelets in ITP could be adequately labelled and total-body scintigraphy performed (Thakur et al, 1976). Using 111In it became evident that there was considerable heterogeneity in platelet turnover between patients, with a substantial proportion having platelet production within normal limits even though the mean cell life of platelets was significantly reduced compared with that in healthy subjects (Stoll et al, 1985; Heyns et al, 1986; Ballem et al, 1987). If platelet destruction were the only mechanism to cause thrombocytopenia, then platelet production would be expected to increase to offset low platelet counts. It was therefore proposed once again that thrombocytopenia may result not only from platelet destruction, but also from antibody-mediated damage to megakaryocytes. Evidence to support this hypothesis has accumulated over time.

Studies in the early 1980s had demonstrated that antibodies against platelet antigens would bind to megakaryocytes as well (Rabellino et al, 1981; Vinci et al, 1984). Even before these observations, McMillan et al (1978) reported that IgG produced by cells (grown in vitro) from the spleens of patients with ITP would bind to megakaryocytes, whereas IgG produced by cells from the spleens of healthy controls did not bind to megakaryocytes. More recent in vitro experiments have further defined the role of autoantibodies in patients with ITP. Two studies in particular, by Chang et al (2003) and McMillan et al (2004) support the view that autoantibodies in ITP suppress megakaryocyte production and maturation and platelet release.

Electron microscopy studies have clarified some aspects of the autoantibody-induced damage in bone marrow megakaryocytes from patients with ITP. Extensive megakaryocytic abnormalities were consistently present in a significant percentage of all stages of ITP megakaryocytes (Stahl et al, 1986; Houwerzijl et al, 2004). The most recent of these studies (Houwerzijl et al, 2004) described features characteristic of nonclassic apoptosis, including mitochondrial swelling with cytoplasmic vacuolization, distention of demarcation membranes, and condensation of nuclear chromatin. Para-apoptotic changes could be induced in megakaryocytes derived from CD34+ cells grown in ITP plasma, suggesting that autoantibodies may initiate the cascade of programmed cell death.

T cells and their role in ITP

T-cell abnormalities in ITP have been consistently reported by one group only since the early 1990s (Semple et al, 1991, 1996; Semple & Freedman, 1994). Given that peripheral blood T cells of patients with chronic ITP could secrete interleukin (IL)-2 on stimulation with platelets, Semple et al (1991) suggested that disorder may be the result of an abnormal T helper (Th) -cell defect that could direct autoreactive B cells to differentiate and secrete IgG autoantibodies. Contributions from other groups have significantly expanded our understanding of the involvement of the various subsets of thymic lymphocytes in the pathogenesis of ITP (Filion et al, 1995; Shimomura et al, 1996; Kuwana et al, 1998, 2001; Fogarty et al, 2003).

Chronic ITP is now considered a Th-1 disease, characterized by the oligoclonal accumulation of Th cells (Shimomura et al, 1996; Fogarty et al, 2003; Stasi et al, 2007). Quantitative or qualitative defects in CD4CD25Foxp3+ regulatory T lymphocytes (Tregs) may play an important role in the loss of self-tolerance in patients with ITP and the persistence of disease (Stasi et al, 2008b; Yu et al, 2008), but whether they are a primary or secondary event in the development of the disease is not yet clear.

Olsson et al (2003) made the intriguing observation that in patients with ITP, devoid of platelet autoantibodies, thrombocytopenia can be mediated by cytotoxic T lymphocytes (Tc). Patients with active disease had peripheral blood Tc that could bind to platelets in vitro and cause significant lysis of the platelets, whereas those patients in remission had little antiplatelet Tc reactivity (Olsson et al, 2003).

The same group also studied the bone marrow and peripheral blood of patients with chronic ITP and suggested that chronic ITP may be a disease of increased T-cell activation due to a Treg defect within the bone marrow. This mechanism may contribute to the autoantibody-mediated suppression of megakaryocyte maturation and platelet production in ITP. In vitro experiments from other investigators have suggested that activated CD8+ T cells in bone marrow of patients with chronic ITP might suppress megakaryocyte apoptosis, leading to impaired platelet production (Li et al, 2007).

Thrombopoietin in ITP

Thrombopoietin (TPO), the primary regulator of platelet production, was purified and cloned in 1994 using various strategies (Bartley et al, 1994; Kuter et al, 1994; Lok et al, 1994; de Sauvage et al, 1994). TPO acts through the TPO receptor (also referred to as c-mpl) to promote megakaryocyte and megakaryocyte precursor proliferation, differentiation, and maturation. TPO levels in ITP are low compared to those with thrombocytopenia from aplastic anaemia, suggesting inappropriate thrombopoiesis in most patients with ITP (Emmons et al, 1996; Kosugi et al, 1996).

This observation has led to one of several new therapeutic approaches to thrombocytopenia in ITP, aimed in this instance at inducing platelet production through growth factor stimulation of megakaryopoiesis (Kuter & Gernsheimer, 2009).

Current view of the pathophysiology of ITP

It is now fair to say that both sides of the 20th century controversy had been partly right. Chronic ITP is a disorder of both increased platelet destruction and inadequate production. In most patients with ITP there is an inadequate bone marrow response to peripheral platelet destruction. In patients with chronic ITP, the megakaryocyte also appears to be a target of antibody and probably of cytotoxic T-lymphocyte injury resulting in ineffective thrombopoiesis. Unlike thrombocytopenia from bone marrow failure syndromes (aplastic anaemia, chemotherapy, myelodysplastic syndromes, etc.) in patients with ITP blood TPO levels are not, or only marginally, increased.

New thinking about the pathophysiology of ITP has been stimulated by recent epidemiological studies challenging the traditional view that ITP is a disease affecting predominantly young women. These studies have shown that ITP becomes more prevalent with increasing age, and that while the incidence of ITP is higher in younger women there is no gender preponderance in older individuals (Frederiksen & Schmidt, 1999; Neylon et al, 2003; Schoonen et al, 2009). The reasons for these changes in the demographics of ITP are not clear, but it has been suggested that they might be related to emerging chronic infections, such as hepatitis C virus (HCV), human immunodeficiency virus (HIV), and Helicobacter pylori (Stasi et al, 2009a). Helicobacter pylori (H. pylori) infection in particular has been recently under intensive clinical investigation. Interestingly, in many countries with a high prevalence of the infection, bacterial eradication reverses the thrombocytopenia in about 50% of cases with chronic ITP. The situation is different in North America, where H. pylori infection is found in a low proportion of cases and eradication seldom has any effect on the platelet count (Stasi et al, 2009b). A plausible explanation is that the H. pylori strains differ, but additional research is required to fully understand this phenomenon.

Changes in the definition of the different phases of ITP

In the past the lack of consensus on standardized critical definitions, outcome criteria, and terminology have been a major obstacle in interpreting the results of clinical studies and in producing reliable meta-analyses of existing data. To address these issues, an International Working Group of recognized experts convened a 2-d structured face-to-face consensus conference in Vicenza, Italy (the Vicenza Consensus Conference) in October 2007, and the report was published in 2009 (Rodeghiero et al, 2009).

Traditionally, ‘acute ITP’ has been used to describe a self-limited form of the disease (e.g. secondary to viral illness in children) whereas ‘chronic ITP’ identified ITP lasting for more than 6 months. In the absence of reliable predictive clinical or laboratory parameters of disease duration, the term ‘newly diagnosed ITP’ is now adopted for all cases at diagnosis. A new category, called ‘persistent ITP,’ was introduced for patients with ITP to define the period lasting between 3 and 12 months from diagnosis. This category includes patients not achieving spontaneous remission or not maintaining their response after stopping treatment between 3 and 12 months from diagnosis. The term ‘chronic ITP’ is used for patients with ITP lasting for more than 12 months (Rodeghiero et al, 2009).

Evolution of management

When Werlhof observed the girl with ITP in 1735, his armamentarium of therapeutic options was obviously very limited. Phlebotomy was commonly prescribed for a number of conditions, but was not considered appropriate in haemorrhagic disorders. Werlhof’s patient apparently recovered with Elixirium acidum halleri (citric acid), although it is more sensible to think of a spontaneous remission of the disease.

Several decades later Willan ascribed the cause of this disorder to ‘a sedentary mode of life, poor diet, impure air and anxiety of mind’, and, as treatment, he recommended ‘moderate exercise in the open air, a generous diet, and the free use of wine, …’ (Willan, 1808). Purgatives have also been used extensively during the 19th century (Woodforde, 1828), and their use was extolled by Eustace Smith (1835–1914) in his book on diseases in children (Smith, 1886).

Since it was first successfully performed in 1916, splenectomy became the mainstay of management in adult patients with refractory, severe ITP. By 1926, Whipple was able to collect data from about 81 patients who had undergone splenectomy for ITP (Whipple, 1926). Two years later Spence collected 23 additional records (Spence, 1928), and in 1932 Eliason and Ferguson brought the collected experience to 213 cases (Eliason & Ferguson, 1932). At that time splenectomy was clearly established as the only definitive therapeutic procedure for ITP. Other than splenectomy, no treatment with a specific rationale was used until the 1950s. Blanchette and Freedman (1998) reported that alternative treatments postulated during this period included irradiation by mercury vapour lamps, administration of snake venom, and irradiation of the spleen.

The use of corticosteroids and adrenocorticotropic hormone (ACTH) has been described since 1951 (Wintrobe et al, 1951). Standard dose prednisolone has been considered the standard initial treatment for newly diagnosed ITP. Immunosuppressive agents were introduced in the 1960s, when the autoimmune nature of ITP was clarified. Over the past 40 years a large number of immunosuppressive drugs have been proposed to be of benefit in patients with ITP, often based on anecdotal reports or uncontrolled series and few are evidence-based (Vesely et al, 2004).

The side effects of treatment are an important consideration and often guide the clinician’s choices. Splenectomy is associated with a number of intra- and peri-operative complications, although the laparoscopic procedure is associated with lower complication and mortality rates than open splenectomy (Kojouri et al, 2004). The risk of overwhelming post-splenectomy is probably small, but quantifying the risk is difficult because of the lack of consistent data. Complications of immunosuppressive therapy are also a major concern and contribute significantly to mortality in ITP (Portielje et al, 2001; McMillan & Durette, 2004).

A milestone in the treatment of symptomatic ITP in children was the introduction of intravenous immunoglobulin (IVIG) by Imbach et al (1981). The efficacy of this treatment was subsequently validated both in adults (Newland et al, 1983) and in pregnancy (Newland et al, 1984). Several mechanisms have been proposed to explain both the acute and long term effects of IVIG, including saturation of Fc receptors on macrophages, upregulation of the inhibitory low affinity IgG receptor FcγIIRB, the regulatory properties of anti-idiotypic antibodies, enhanced suppressor T lymphocyte function and decreased autoantibody production (Lazarus & Crow, 2003). The use of anti-D immunoglobulin was based on studies by Abdulgabar Salama and colleagues indicating that the direct antiglobulin test was often transiently positive after IVIG and other laboratory parameters (e.g. bilirubin and haptoglobin) were also consistent with haemolysis (Salama et al, 1983). These investigators documented platelet responses in Rhesus-positive ITP patients following the intravenous administration of an anti-D product and introduced the concept of macrophage blockade (Salama et al, 1983). James Bussel and his group later expanded the knowledge about the modalities of treatment with anti-D in various settings (Bussel et al, 1991; Scaradavou et al, 1997; Newman et al, 2001; Cooper et al, 2002; Michel et al, 2003; Kane et al, 2010).

With advances in molecular biology and pharmacological technologies, targeted therapy has been investigated since the 1980s. Early experience with a monoclonal antibody directed against the Fc gamma-receptor in a patient with refractory ITP resulted in a transient increase in platelet counts (Clarkson et al, 1986). Moderate enthusiasm was also generated by targeting the CD52 antigen with alemtuzumab (Lim et al, 1993; Willis et al, 2001), and more recently by clinical trials with anti-CD40 ligand (CD154) antibodies (Patel et al, 2008). However, the most consistent results with monoclonal antibody therapy have been obtained with rituximab, an anti-CD20 chimeric antibody inducing B cell depletion. At the dawn of the new millennium rituximab was reported an efficacious therapy for a significant proportion of adults with chronic ITP (Saleh et al, 2000; Stasi et al, 2001) and this agent soon became the standard (albeit unlicensed) treatment for patients with this condition in many countries.

There is no doubt that the major breakthrough in the treatment of chronic ITP has been witnessed in the last few years, with the publication of the results of randomized clinical trials with the TPO receptor agonists, more specifically romiplostim (Kuter et al, 2008, 2010) and eltrombopag (Bussel et al, 2009a; Cheng et al, 2010). Romiplostim is a an Fc-peptide fusion protein (or ‘peptibody’) that binds to the extracellular domains of the TPO receptor. However, it has no sequence homology to endogenous TPO and theoretically avoids the risk of eliciting cross-reacting, neutralizing antibodies to TPO. Eltrombopag is an orally administered, small molecule non-peptide that selectively binds to the transmembrane domain of the TPO receptor and is metabolized by the liver. Systemic exposure to the drug may be increased in individuals of East-Asian descent; as a consequence, an initial dose decrease to 25 mg/d (rather than the 50 mg/d used for other ethnicities) is recommended in these patients (Garnock-Jones & Keam, 2009).

Both romiplostim and eltrombopag have shown response rates unequalled by previous medical therapies and are almost as efficacious in splenectomized patients as in the nonsplenectomized ones. Recent data have confirmed the efficacy and safety following long term usage of both currently available products (Bussel et al, 2009b; Saleh et al, 2009).

Ongoing clinical trials in ITP involve antibodies against the Fc receptor such as MDX-33, a humanized anti-FcγRI monoclonal, and GMA-161, a humanized anti-FcγRIII monoclonal (Li & Hou, 2008). Investigation of inhibition of FcR signalling mechanisms is currently under investigation with R788 (R935788), a small molecule prodrug of the biologically active R406 (Podolanczuk et al, 2009). This is a potent and relatively selective orally available inhibitor of Syk (spleenz tyrosine kinase).

What the future holds

Since the original description by Werlhof, much has been learned about the pathophysiology and management of ITP. It is now understood that, in ITP, both the B-cell and the T cell compartments are deranged, and that the thrombocytopenia is the result of both depressed platelet production and increased platelet destruction. Treatment has evolved from splenectomy and nonspecific immunosuppression to stimulation of platelet production with thrombopoietic agents and targeted therapies. We have currently in our hands a relatively wide armamentarium of effective therapeutic options.

However, many important issues still need to be adequately addressed in future research. Major gaps remain the lack of positive diagnostic criteria as well as of clinical tools to guide management. Analysis of the single nucleotide polymorphisms (Suzuki et al, 2008; Satoh et al, 2009; Zhao et al, 2010) and of platelet proteome (Qureshi et al, 2009) might provide insightful data.

Two other important issues are fatigue and thrombophilia. Abnormal tiredness affects a substantial number of patients with ITP (Sarpatwari et al, 2010a), but the severity of this symptom does not fully correlate with the platelet count or bleeding manifestations (McMillan et al, 2008). The mechanisms of fatigue are poorly understood and this is an area that needs further detailed investigation. Whether successful treatment of ITP can improve fatigue has been recognized (Mathias et al, 2008) but not fully demonstrated. Results from two recently published randomized trials with the new TPO receptor agonists are actually contradictory. Compared to placebo or standard of care, a positive effect on fatigue has been reported following 6 months of treatment with eltrombopag (Cheng et al, 2010) but not after 6 or 12 months of treatment with romiplostim (Kuter et al, 2008, 2010).

Recent evidence from a retrospective analysis also suggests that ITP is a pro-thrombotic condition (Sarpatwari et al, 2010b). While a confirmatory prospective study is needed, the results of this study suggest that ITP patients with supra-normal responses to treatment may require anti-platelet therapy, such as low dose aspirin, and that limiting the target platelet count following therapy is sensible. It is clear, however, that even those patients who remain thrombocytopenic are at potential risk of increased arterial and venous thrombo-embolism.

With regard to treatment, there are a variety of other potential approaches. One such approach may be blockade of B-cell survival signals. B cell-activating factor/B lymphocyte stimulator (BAFF/BLyS) and a proliferation-inducing ligand (APRIL), and their receptors consisting of BAFF receptor (BAFF-R)/BLyS receptor 3 (BR3), trans membrane activator and calcium modulator and cyclophilin ligand interactor (TACI), and B cell maturation antigen (BCMA), constitute a well-characterized system involved in B-cell development and survival. Dysregulation of this system may be involved in the development and/or maintenance of autoimmunity in ITP (Emmerich et al, 2007). On the T cell side, direct interference with the profile and/or function of regulatory T cells (Tregs) may also be applied clinically to reverse the abnormal pattern of cellular immunity. Patients with chronic ITP are characterized by a decreased number and activity of Tregs (Stasi et al, 2008b; Yu et al, 2008). In vitro data suggest that modulation of Treg activity can be achieved by exposure to cytokines such as IL-2 and IL-6, and suppressor of cytokine signalling 3 (SOCS3) may be a potential therapeutic target (Pillemer et al, 2007). Also, adoptive therapy using Tregs of the correct specificity in sufficient numbers can be a potent tolerogenic regimen in combination with other therapies that target the pathogenic T cells.

Finally, combinations of agents with different mechanisms of action and molecular targets may be a rational approach, but the schedule and dosing will require extensive investigation in clinical trials.

Currently, although there is general agreement over treatment of the newly presenting adult with ITP, much systematic study is required. While there is an increasing understanding that the patient should be treated for their clinical state rather than their platelet count, there is little agreement on the second line treatment in the relapsed or refractory patient or of the true natural history of the disease. To this end an international group produced a consensus report on investigation and management, giving (where possible) evidence-based advice on treatment pathways (Provan et al, 2010). It is hoped that by following such an approach treatment in the future can be audited and authoritative guidelines developed.